Art Doc Viz – Interactive Visualization of Artist Documentation

Over the past two decades, digital archiving and building interfaces to access digital archives has been a major challenge at the intersection of computer sciences, arts and humanities. In our project, we approach social media data and website data from artists as a way of self-documentation that we want to archive and bring in context. In this text, we outline the making of Art Doc Viz, an interactive visualization that builds on an open source and machine-learning-based data pipeline designed to bring arbitrary amounts of artists' self-documented work from different sources (for example Instagram and websites) into a visualization structure that makes it accessible interactively and in 3D, and that makes relations between the artists' documentations visible. This is achieved by using a series of AI tools like entity recognition and image recognition and UMAP as a clustering algorithm. Finally, the visualization uses advanced 3D graphics to enable the user to explore the visualization in an intuitive way.